ICVS 2011
8th International Conference on Computer Vision Systems
20 - 22 September 2011, Sophia Antipolis (France)

logo Inria
with the support of
logo region
logo casa
PROGRAM (in progress)

All the sessions will be held in the Jacques Morgenstern amphitheatre, Gilles Kahn Building.

Monday 19
Workshop on Health & Well Being
Tuesday 20
ICVS conference
Wednesday 21
ICVS conference
Thursday 22
ICVS Conference
Friday 23
Workshop on Behaviour Analysis and Video Understanding
8:30 - 9:00
      9:00 - 9:10 Welcome
9:00 - 9:30
9:00 - 10:00
Invited talk
G. Bradski
9:00 - 10:00
Invited talk
M. Shah

9:10 - 9:50 Invited talk M. Mokhtari

9:30 - 10:30
Session SmartHome 1
10:00 - 10:30
Coffee break
10:00 - 10:30
Coffee break
9:50 - 10:50
Session 1
10:30 - 11:00
Coffee break or demos
10:30 - 12:00
Performance Evaluation 1
10:30 - 12:00
Control of Perception 1
10:50 - 11:10
11:00 - 12:00
Session SmartHome 2
12:00 - 13:30
12:00 - 14:00
11:10 - 12:40
Session 2
12:00 - 13:00
Session Demo
12:45 Chek-in opening 12:40 - 14:00
13:00 - 14:15
Lunch and Demos
14:00 - 14:30
Welcome and ICVS Opening
13:30 - 14:30
Performance Evaluation 2
14:00 - 15:00
Knowledge Directed Vision
14:00 - 14:40
Invited talk
R. Chang
14:15 - 15
14:30 - 15:30
Invited talk J. Stam
14:30 - 15:00
Coffee break
15:00 - 15:30
Coffee break
14:40 - 15:40
Session 3
15:15 - 16:15
Robotics for H & WB
15:30 - 16:00
Coffee break
15:00 - 17:00
Activity Recognition
15:30 - 17:30
Control of Perception 2
15:40 - 17:15
16:15 - 16:45
Coffee break or Demos
16:00 - 18:00
Vision Systems
17:00 - 22:30
Social Event
16:45 - 18:15
Social Interaction and Robotics
18:00 - 19:00
Welcome Cocktail
Tuesday, September 20, 2011

14:00 - 14:15  Welcome at Inria
Gérard Giraudon, Director of the Inria Sophia Antipolis - Méditerranée research centre
14:15 - 14:30  Conference Opening
Monique Thonnat, Conference Chair
James Crowley, Program Chair
Bruce Draper, Program Chair
14:30 - 15:30
Invited Talk (I): Vision everywhere: An overview of vision on mobile consumer device and popular PC hardware, Joe Stam, NVIDIA Corp., Santa Clara, CA, USA.:
Session Chair(s): J. Crowley
For the past couple decades computer and machine vision technology served largely in scientific and industrial applications such as factory automation or medical imaging. Many factors limited wide adoption, including the prohibitive cost of sufficient processing power and availability of suitable cameras.
The last several years have ushered in a huge surge in the use of vision for mainstream consumer applications. Vision had become the latest craze for video game consoles, and with over a billion camera phones sold every year powerful vision capable devices will be in everyone's pocket. In this session, we'll provide an overview of the latest parallel processors for computer vision, focusing particularly on NVIDIA's Tegra processors and GPU computing architectures. We'll also highlight the latest development tools and libraries for mobile vision application development.
15:30 - 16:00 coffee break
Vision Systems:
Session Chair(s): M. Vincze
Knowing what happened - Automatic Documentation of Image Analysis Processes
Authors:  Birgit Moeller (University Halle-Wittenberg) , Oliver Gress (University Halle-Wittenberg) , Stefan Posch (University Halle-Wittenberg)

Efficient Use of Geometric Constraints for Sliding-Window Object Detection in Video
Authors:  Patrick Sudowe (RWTH Aachen) , Bastian Leibe (UMIC Research Centre, RWTH Aachen)

A method for asteroids 3D surface reconstruction from close approach distances
Authors:  Luca Baglivo (University of Trento) , Alessio Del Bue (IIT) , Lunardelli Massimo (University of Trento) , Francesco Setti (University of Trento) , Vittorio Murino (IIT) , Mariolino De Cecco (University of Trento)

RT-SLAM: a generic and real-time visual SLAM implementation
Authors:  Cyril Roussillon (LAAS/CNRS) , Aurelien Gonzalez (LAAS/CNRS) , Joan Sola (LAAS/CNRS) , Jean-Marie Codol (LAAS/CNRS) , Nicolas Mansard (LAAS/CNRS) , Simon Lacroix (LAAS/CNRS) , Michel Devy (LAAS/CNRS)

18:00 - 19:00 Welcome Cocktail
Wednesday, September 21, 2011
09:00 - 10:00
Invited Talk (II): Creating perception systems using OpenCV, Gary Bradski, Senior Scientist, Willow Garage Inc., Menlo Park, CA, USA
Session Chair(s): B. Draper
Abstract :
The Open Source Computer Vision Library (OpenCV) is now used extensively around the world, approaching 3M downloads with an active user group of over 45,000 people. OpenCV is getting development support from Willow Garage, NVidia and Google as well as wide spread community contributions. Some of the main environments that OpenCV supports are: Linux, Windows, Mac OS, Android and iPhone ... along with a host of robots through ROS, the Robot Operating System.
This talk will overview OpenCV: where it is and where it is going, including a program to standardize a slimmed down core subset of OpenCV into the Khronos standard to allow hardware manufactures to better build optimize vision hardware around it. I will highlight some recent functions and then demonstrate a new development tool that allows a programmer to code mostly in Python but have the runtime system run entirely in C++ also guaranteeing synchronization for 2D and 3D data. This environment will further allow the developer to package their functionality as a standalone module or as a node in ROS and to seamlessly store and retrieve data, models and parameters to and from a database. We are using this system to develop object recognition and SLAM applications for robotics. I will demonstrate some of this functionality.
10:00 - 10:30 coffee break
10:30 - 12:00
Performance Evaluation 1:
Session Chair(s): B. Draper
A quantitative Comparison of speed and reliability for the log-polar mapping techniques
Authors:  Manuela Chessa (University of Genoa) , Silvio Sabatini (University of Genoa) , Fabio Solari (University of Genoa) , Fabio Tatti (University of Genoa)

Toward accurate feature detectors performance evaluation
Authors:  Pavel Smirnov (Intel Labs) , Piotr Semenov (Intel Labs) , Alexander Redkin (Intel Labs) , Anthony Chun (Intel Labs)

Evaluation of local descriptors for action recognition in videos
Authors:  Piotr Bilinski (Inria) , Francois Bremond (Inria)

12:00 - 13:30 lunch
13:30 - 14:30
Performance Evaluation 2:
Session Chair(s): F. Brémond
On the Spatial Extents of SIFT Descriptors for Visual Concept Detection
Authors:  Markus Mühling (University of Marburg) , Ralph Ewerth (University of Marburg) , Bernd Freisleben (University of Marburg)

An Experimental Framework for Evaluating PTZ Tracking Algorithms
Authors:  Pietro Salvagnini (IIT) , Marco Cristani (IIT) , Alessio Del Bue (IIT) , Vittorio Murino (IIT)

14:30 - 15:00 coffee break
15:00 - 17:00
Activity Recognition:
Session Chair(s): J. Crowley
Unsupervised activity extraction on long-term video recordings employing soft computing relations
Authors:  Luis Patino (Inria) , Murray Evans (Reading University) , James Ferryman (Reading University) , Francois Bremond (Inria) , Monique Thonnat (Inria)

Unsupervised Discovery, Modeling and Analysis of long term Activities
Authors:  Guido Pusiol (Inria) , Francois Bremond (Inria) , Monique Thonnat (Inria)

Ontology-based Realtime Activity Monitoring Using Beam Search
Authors:  Wilfried Bohlken (University of Hamburg), Bernd Neumann (University of Hamburg), Lothar Hotz (HITeC), Patrick Koopmann (University of Hamburg)

Probabilistic Recognition of Complex Event
Authors:  Rim Romdhane (Inria), Bernard Boulay (Inria), Francois Bremond (Inria), Monique Thonnat (Inria)

17:00 - 22:30 Social event (guided tour and gala dinner in Antibes)
Thursday, September 22, 2011
09:00 - 10:00
Invited Talk (III): Video surveillance systems - UCF computer vision lab experience, Prof. Mubarak Shah, University of Central Florida, Orlando, FL, USA
Session Chair(s): M. Thonnat
Video surveillance and monitoring is one of most active areas of research in Computer Vision. The main steps in a video surveillance system include: detection and categorization of objects of interest in video (e.g. people, vehicles), tracking of those objects from frame to frame, and recognition of their activities, behavior and patterns. Illumination changes, shadows, occlusion and clutter etc make moving object detection difficult. Dealing with occlusion and object interaction in high density scenes, and tracking across multiple overlapping and non-overlapping field cameras are some of the challenges in tracking. Activity recognition is an area of active research and learning based solutions seem to be the most promising ones and there are still many challenges to be addressed. Over the last decade a number of prototype Surveillance Systems have been developed at UCF Computer vision lab, employing in-house developed point solutions for above components.
In this talk I will first present key techniques in object detection, tracking and activity detection, then discuss the software engineering aspects of two of those systems namely: KNIGHT: Ground Surveillance System for low density scenes, and COCOA - a solution for tracking moving objects in aerial videos. I will discuss the lifecycle of systems developed in a typical academic environment, the typical code quality issues encountered and efforts being undertaken to improve continuously.
10:00 - 10:30 coffee break
10:30 - 12:00
Control of Perception (I):
Session Chair(s): M. Thonnat
Learning what Matters: Combining Probabilistic Models of 2D and 3D Saliency Cues
Authors:  Ekaterina Potapova (TUW) , Michael Zillich (TUW) , Markus Vincze (Vienna University of Technology)

3D Saliency for Abnormal Motion Selection: the Role of the Depth Map
Authors:  Nicolas Riche (University of Mons), Matei Mancas (University of Mons), Nicolas Riche (University of Mons), Bernard Gosselin (University of Mons), Thierry Dutoit (University of Mons)

Scene Understanding through Autonomous Interactive Perception
Authors:  Niklas Bergström (KTH) , Carl Henrik Ek (Royal Institute of Technology) , Mårten Björkman (Royal Institute of Technology) , Danica Kragic (KTH, Sweden)

12:00 - 14:00 lunch
14:00 - 15:00
Knowledge Directed Vision:
Session Chair(s): B. Neumann
A Cognitive Vision System for Nuclear Fusion Device Monitoring
Authors:  Vincent Martin (CEA) , Victor Moncada (CEA, IRFM) , Jean-Marcel Travere (CEA, IRFM) , Thierry Loarer (CEA) , Guillaume Charpiat (Inria) , Francois Bremond (Inria) , Monique Thonnat (Inria)

Knowledge Representation and Inference for Grasp Affordances
Authors:  Karthik Mahesh Varadarajan (TU Vienna), Markus Vincze (TU Vienna)

15:00 - 15:30 lunch
15:30 - 17:30
Control of Perception (II):
Session Chair(s): V. Martin
Towards a General Abstraction Through Sequences of Conceptual Operations
Authors:  Gregor Miller (University of British Columbia) , Steve Oldridge (University of British Columbia) , Sidney Fels (University of British Columbia)

Girgit: A Dynamically Adaptive Vision System for Scene Understanding
Authors:  Leonardo Rocha (Inria) , Sagar Sen (Inria), Sabine Moisan (Inria), Jean-Paul Rigault (Université de Nice - Sophia Antipolis)

Run Time Adaptation of Video-Surveillance Systems: A software Modeling Approach
Authors:  Sabine Moisan (Inria ) , Jean-Paul Rigault (Université de Nice - Sophia Antipolis) , Mathieu Acher (I3S) , Philippe Collet (I3S) , Philippe Lahire (I3S)

Automatically Searching for Optimal Parameter Settings Using a Genetic Algorithm
Authors:  David Bolme (Colorado State University), J. Ross Beveridge (Colorado State University), Bruce Draper (Colorado State University), Jonathon Phillips (National Institute of Standards and Technology), Yui Man Lui (Colorado State University)


last update : 09/16/2011 - webmaster - credits